Our threats are: PHI (Protected Health Information) potentially being stolen and used to blackmail the corresponding gov bodies and persons by either a man in the middle or an internal malicious cloud provider employee.
Our goal is: protecting the PHI from the security flaws in the public cloud, ensure the CIA triad (confidentiality is enabled thru encryption, integrity thru blockchain, availability via IPFS)
Our challenges are: securing the data properly in the public cloud, so that no bad actors from the inside or the outside can access them.
We have the following IT security architecture plan:
where M/TEXT is the proprietary document server, MS MA stands for the potentially malicious privacy-disrespecting cloud provider employee, Kunden IT-System is the on-premise customer IT architecture, Editor is a HTML5 App for editing the documents/templates itself, M/Workbench is proprietary document editor software, the VPN arrow thingy is not a Plex logo and represents a VPN tunnel, XML contains the PII and PHI, IPLD connects the Git, Blockchain and IPFS for using in the proprietary document editor software, the weird contraption around the PostgresSQL DB represents a proprietary encryption method (I know that it's bad to use proprietary encryption, but it's a given)
Let's assume we have our metadata for PHI stored as hashes in the blockchain, the actual PHI in IPFS. This is realized thru two connected smart contracts as shown in the scheme, where one only contains the "pseudo UID" and PII-hash without the PII itself and the other actually stores the IPFS URL for the PII and the PII itself.
Our questions are:
- Is this architecture good enough for our use case?
- Are there any general flaws in the Cybersecurity architecture?
- Is it smart to use Blockchain and IPFS here?
- Does it bring any flaws by itself?
